> On Mon, 2010-07-26 at 12:49 -0400, John Martin wrote:
> > Slight hijack of this thread, but somewhat on
> topic.
> > The man page for ddi_fls() says:
> >
> > int ddi_fls(long mask);
> > ...
> > mask A 32-bit argument value to search
> through.
> >
> > I assume the man page was written prior to LP64.
> > What is the expected behavior on LP64?
>
> I'd expect that on LP64 its just an int rather than a
> long, but you can
> check the code yourself. When last I looked,the code
> for this function
> was quite easy to grok.
There is only one function for ILP32 and LP64 and it uses
long for the mask, matching the man page. The code
lifted from common/os/sunddi.c:
$ cat fls.c
#include <stdlib.h>
#include <stdio.h>
#include <strings.h>
int
ddi_fls(long mask)
{
while (mask) {
long nx;
if ((nx = (mask & (mask - 1))) == 0)
break;
mask = nx;
}
return (ffs(mask));
}
int
main(int argc, char *argv[], char *envp[])
{
long val = -1;
printf("fls(%ld/0x%lx) = %d\n", val, val, ddi_fls(val));
return (0);
}
$ cc -m32 fls.c
$ ./a.out
fls(-1/0xffffffff) = 32
$ cc -m64 fls.c
$ ./a.out
fls(-1/0xffffffffffffffff) = 0
$
--
This message posted from opensolaris.org
_______________________________________________
driver-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/driver-discuss
